Neurocomputing

The primary objective of system neuroscience is to understand the functional mapping and its causation in the dynamics of the brain network. Some experimental and methodological studies suggest that functional modularity and its hierarchical information processing in the brain network are crucial to understanding the functional role of task-specific or state-specific information flow in the brain. However, because most of the established techniques for detecting effective network structures in the neuroscience research field are strongly based on the "Granger causality" perspective, existing causal discovery methods specified for brain network analysis cannot identify the causal hierarchy in the modular network in the brain due to spurious correlation issues and indistinguishability of causal direction under the Gaussianity of observational noise in a linear system. To address the issues, we developed a causal discovery method for synchronous neural dynamics, called the Jacobian-informed linear non-Gaussian acyclic model, "j-VAR-LiNGAM", by incorporating the information of the Jacobian matrix determined from a phase-coupled oscillator model estimated from observed neural data into the VAR-LiNGAM algorithms. The method was validated by showing that it could extract causal ordering in both synthetic data and empirical neural observed data. Moreover, by analyzing the observed neural oscillatory signals obtained from mice and humans, we confirmed that our method identified causally hierarchical structures in the brain, which aligned with the neurophysiological interpretations. These findings suggested that our proposed method can reveal the neural basis of hierarchical information processing in the brain network.

Wave recurrent neural networks (wRNNs) are biologically inspired recurrent architectures that use traveling-wave dynamics to support sequence learning and memory. However, their input-to-hidden pathway remains relatively simple compared with biological neurons, where dendrites perform nonlinear input integration. In this study, we introduce the Dendritic Wave Recurrent Neural Network (DWRNN), which augments the input pathway of the wRNN with nonlinear basal dendritic branches while preserving the original recurrent wave dynamics. We evaluate DW-RNN on a simple copy task, sequential MNIST (sMNIST), permuted sequential MNIST (psMNIST), and noisy sequential CIFAR-10 (nsCIFAR-10). On the copy task, DW-RNN shows learning behavior comparable to the standard wRNN, suggesting that dendritic input integration does not disrupt the recurrent wave-based memory mechanism. On the three sequential image-classification benchmarks, DW-RNN outperforms the standard wRNN, improving accuracy from 97.27 {+/-} 0.15% to 97.82 {+/-} 0.12% on sMNIST, from 96.74 {+/-} 0.17% to 96.92 {+/-} 0.10% on psMNIST, and from 54.30 {+/-} 0.79% to 55.65 {+/-} 0.55% on nsCIFAR-10. In addition to improving mean accuracy, DW-RNN exhibits lower across-seed variability on all three classification benchmarks, suggesting that dendritic input integration may improve the stability of wRNN training. Hidden-activity visualizations further show that DW-RNN preserves the characteristic traveling-wave patterns of the original wRNN. These results suggest that dendritic computation and traveling-wave recurrent dynamics provide complementary mechanisms for biologically inspired sequence learning.

Stabilizing synaptic plasticity together with the neurons activity has remained a central challenge in theoretical neuroscience since the introduction of Hebbian learning principles. Classical Hebbian learning rules typically lead to unbounded synaptic growth, motivating the development of stabilization mechanisms such as normalization methods, BCM-type learning, synaptic scaling and others. While these approaches can prevent divergence, they can also exhibit different limitations e.g. resulting in too-sparse synaptic configurations or leading to poor scalability with increasing network size. A recently introduced meta-plasticity mechanism, termed annealed linear learning (ALL), dynamically reduces the learning rate as neuronal output increases, thereby preserving stable and interpretable fixed-point behavior of the output. However, the original formulation leads to an irreversible decay of the learning rate, preventing adaptation to changing environmental conditions. To address this, in the present study, we balance learning rate reduction at large outputs with recovery at small outputs and in addition introduce forgetting that gradually reduces synaptic weights. These extensions allow the system to discard outdated representations and adapt to novel input conditions. Analytical investigations demonstrate that the favorable output fixed-point properties of the original ALL framework are preserved under the extended rule. Furthermore, simulations with an artificial agent show that the proposed mechanism enables robust and fast re-learning and adaptation in changing environments.

Abnormal states such as erroneous behaviors are generally difficult to represent from neural data. However, such states are also known to have specific spatiotemporal features, indicating a feasibility of developing a method to focus on them. If a method can highlight these spatiotemporal features, it may effectively represent such abnormal states, helping evaluate abnormal brain functions. In the present study, we proposed the hierarchy of supported modules (HSM) to highlight spatiotemporal features that can represent abnormal states. HSM spatiotemporally transforms multidimensional neural time-series based on their spatiotemporal context. We evaluated HSM through decoding and similarity analyses using multiple publicly available datasets. In the HSM results, decoding accuracies were higher for erroneous behaviors than for normal behaviors, and similarities were lower between erroneous behaviors and normal behaviors than between normal behaviors, demonstrating the ability of HSM to capture the spatiotemporal features of erroneous behaviors. Surprisingly, many parts of these results were also present even before HSM learning, showing the virtue of HSM as a simple-to-use method. The proposed HSM method may help elucidate the mechanisms underlying erroneous behaviors.

Equilibrium propagation (EP) is a biologically plausible alternative to backpropagation that has demonstrated competitive performance across a range of machine learning tasks. Recent work has extended EP to spiking neural networks (SNNs), leveraging leaky integrate-and-fire (LIF) neurons and spike-based plasticity rules to improve biological realism while maintaining strong performance. In this work, we propose an EP-based SNN framework that combines LIF neural dynamics with a predictive learning rule, replacing conventional spike-timing-dependent plasticity (STDP) with a learning rule more directly aligned with predictive coding principles. We evaluate the proposed model on multiple image classification benchmarks, including MNIST, KMNIST, and Fashion-MNIST, and compare its performance with a BP-trained LIF SNN baseline. Our results show that the proposed EP-based LIF model (EP+LIF) achieves competitive accuracy across datasets, with performance approaching that of the BP-trained counterpart (BP+LIF) while relying on a biologically motivated local learning rule. In addition, analysis of hidden-layer spiking activity reveals that EP+LIF produces more persistent hidden-state activity, whereas BP+LIF yields sparser spiking representations. These results demonstrate that predictive learning can support effective EP-based training in LIF spiking networks, while also highlighting differences in activity patterns that motivate future work on activity regulation and sparse spiking dynamics.

The organizational principles of natural neural networks could inspire the new architecture design of artificial neural networks (ANNs). Analysis of single-neuron connectomes of mouse brains revealed distinct profiles of three-node connectivity motifs in various cortical areas and hippocampal formation. A connectome-informed neural network algorithm ("CINA") was developed to incorporate natural connectivity motifs into ANN algorithms represented by recurrent neural network (RNN) and transformer-based large language model (LLM). We found that incorporation of the average profile of cortical motifs improved the RNNs performance in noise-resistant categorization and motor learning benchmark tasks, as compared with RNNs with random connectivity. Notably, incorporating cortex-specific motifs further elevated the RNNs performance in tasks related to the cortical function, and this effect was enhanced by artificially increasing the bias in the motif profile. Similar experimental results were verified on an LLM using Motif-Transformer for natural language question answering and brain-signal decoding tasks. Graph-theoretic analyses showed that incorporating natural motifs drove the emergence of modular and small-world properties in ANNs. Together, we demonstrated not only connectome-inspired optimization of ANN architecture but also functional significance of specific motif profiles in various cortices.

BackgroundThis study examines a competition-based model (C-model) designed to capture the temporal dynamics of successive brain microstates derived from electroencephalography (EEG) recordings during eyes-open conditions. The analyzed data were obtained from a public repository comprising microstate sequences from 60 sessions of a single subject [1]. When applied to microstate dynamics, the C-model posits a stochastic competition among neural circuits underlying the expression of individual microstates. MethodsThe model is formulated at a conceptual level (computational level in Marrs framework) and employs a geometric distribution to account for the long right tail of microstate duration distributions, interpreted as the probability of "failure" of the currently active microstate to persist. To account for the short-lived left tail, the model incorporates a transient increase in the stability of the currently active network, or equivalently, a temporary decrease in the activation probability of competing microstates (refractory period). ResultsThe model provides a good fit to the microstate duration distributions across all 60 sessions. One third of sessions showed microstate identity sequential dependency with respect to the previous microstates. DiscussionThese results suggest that the C-model captures key aspects of microstate temporal structure. Moreover, because microstate probabilities can be modulated by psychophysiological conditions--including the influence of previously active networks--the model may serve as a building block for more comprehensive neurobiological frameworks of neural and behavioral dynamics. In such frameworks, microstate sequences could emerge from structured competition and flow among neural networks supporting microstate expression.

We adapted an olfactory neuromorphic algorithm to image and sound recognition. To achieve this, we carried out specific preprocessing procedures that were tailored to each modality. For images, we used the NIST digits dataset directly. For sound, we used samples from the Google Speech Command dataset. A gammatone filter was applied to each to reduce the noise of the short audio sample and convert the temporal sound signal to a positional frequency signal. The single stimulus test algorithm was then modified to handle audio processing on extracted columns from a gammatone filter spectrogram obtained from the sound file. We also implemented PCA for all modalities, retaining around 90% of the variance. The results showed that over sequential 'olfactory' gamma cycles, the algorithm successfully achieved one-shot online learning over the image and sound modalities as well. However, PCA representations did not attain high similarities to their corresponding templates for all three modalities.

BackgroundThe relevance of spontaneous activity has been unlocked thanks to recent large scale recordings that revealed, via Shared Variance Component Analysis (SVCA), the high-dimensional nature of the ongoing activity. A fundamental problem is how the dimension modifies when more neurons are included in the analysis. Contradictory results have been reported on this subject based on SVCA and Principal Component Analysis (PCA). New MethodWe investigate pro et contra of SVCA and PCA for the identification of reliable responses encoding underlying state variables. We focus on common features of the spectra of the reliable variances (RVs) and on their dimensionality. The analysis is demonstrated on previously published Ca2+ data from the visual and the dorsal cortex in head fixed mice during spontaneous behavior. ResultsRVs grow proportionally to the number N of neurons and show a power-law decay k- with the k-th SVC dimension over a range bounded by a maximal dimension kc, initially diverging as N 1/ and then saturating at sufficiently large N. The reliable dimensionality, estimated with different methodologies, also shows a clear saturation to an asymptotic value for large N. Furthermore, its value decreases when becomes larger, as demonstrated by employing experimental data as well as theoretical predictions. ConclusionWe have shown that SVCA is an extremely effective tool to extract reliable features from the neural signals, and that the exponent represents a biomarker able to reveal the level of correlation of the neurons as well as the dimensionality of the reliable space. HighlightsO_LIAdvantages and drawbacks of Shared Variance Component Analysis to extract reliable signals from neural data C_LIO_LIComparison of different methods to estimate reliable neural dimensionality associated to spontaneous activity C_LIO_LIAnalytical expressions of embedding dimensionality for power-law decaying reliable variances C_LIO_LIBounded growth of the dimensionality with the number of neurons C_LI

Transcranial Electric stimulation (tES) is a safe and noninvasive technique increasingly used in treating brain disorders. Despite many studies on tES, there is still a lack of understanding of its mechanisms at the microscale network level. This is crucial for optimized parameter selection in therapy approaches such as the treatment of pharmacoresistant epilepsy. In this study, we made use of a recently published neuroinspired microscale model of the neocortex, known as NeoCoMM, and integrated a "Lambda E"-based model of tES. This updated version was used to investigate the acute effects of tES (tDCS and tACS), on the neural activity of various neuron types in both healthy and epileptic brain states. Results showed that in the case of healthy alpha and gamma rhythms, tACS induced electric field entrainment at the peak power frequency of the network oscillations as measured by the Local field Potentials (LFPs). This resonance-like entrainment was independent from the individual firing rate of cell types. For epileptic activity, tACS did not provide consistent results. Cathodal tDCS resulted in a promising decrease in hyperexcitable activity throughout simulations. These results advance our understanding of the impact of tES on network dynamics at both the extracellular and intracellular activity levels. Author summary

In patients with epilepsy, seizures are associated with pathological neural synchronization. However, the preictal period preceding a seizure often exhibits reduced spatial synchronization compared to normal cognition. This observation aligns with the concept of the brain as a complex dynamical system, where a reduction in dimensionality and resilience can precede a phase transition. The Critical Brain Hypothesis suggests a connection between the loss of healthy scale-free behavior and various disorders, including epilepsy. Our study investigates preictal changes by utilizing network features, such as mean node degree and mean clustering coefficient, derived from thresholded correlation matrices of patient intracranial electrocorticographic electrode data. We observed a suppression of intermittent high-synchronization periods within the feature space during the minutes leading up to seizure onset. This constriction of the explored hypervolume in the preictal state indicates a breakdown in the brains ability to maintain normal coherence. We use these preictal changes to predict the probability of seizure onset using a Support Vector Machine algorithm. These discrete predictions can then be combined into real-time continuous seizure risk forecasts via Bayesian updating. This innovative and computationally lightweight approach has the potential to significantly improve upon static predictions, providing opportunities for more adaptable, quantitative, and interpretable tools for managing seizures.

Mental health challenges among university students have increased due to academic pressure, lifestyle changes, and continuous digital engagement. Existing approaches for mental health assessment often rely either on self-reported psychological scales or isolated behavioral indicators, limiting their ability to capture complex temporal and contextual patterns. This study proposes an interpretable multimodal framework for student mental health risk assessment using behavioral sensing, academic information, ecological momentary assessments (EMA), and psychometric survey data. A bidirectional Long Short-Term Memory autoencoder is employed to learn latent temporal representations from day-level behavioral sequences, while graph embeddings capture structural relationships among students using similarity-based neighborhood graphs. These representations are fused with academic and survey-derived features and reduced using Principal Component Analysis and Uniform Manifold Approximation and Projection. K-means clustering is then applied to identify behaviorally distinct student groups. Experimental analysis on the StudentLife dataset demonstrates meaningful clustering performance with a Silhouette Score of 0.4209 and Adjusted Rand Index stability of 0.6869. The identified clusters correspond to low-risk, moderate-risk, and high-risk behavioral profiles. To improve interpretability and practical usability, a fuzzy inference system is introduced to compute mental risk, academic risk, and wellbeing indices using psychometric indicators including PHQ-9, PSS, PANAS, VR-12, and Big Five personality traits. The results demonstrate the potential of combining multimodal behavioral modeling with interpretable fuzzy reasoning to support early mental health risk assessment in educational settings.

Mapping hierarchical brain networks within traditional Euclidean space causes significant structural distortion, undermining neuroimaging diagnostic frameworks. While hyperbolic models like the Poincare ball preserve these nested topologies, they demand heavy computational overhead due to intricate Mobius operations and curved geodesics. This paper introduces a highly efficient non-Euclidean framework for analyzing neurocognitive decline utilizing the Beltrami-Klein ball model. By projecting hyperbolic geodesics as Euclidean straight lines, this approach converts complex distance calculations into simple dot products, radically reducing processing demands. We validated our methodology against state-of-the-art Poincare and Lorentz baselines using datasets for Schizophrenia, Parkinsons Disease, and Alzheimers Disease. The Klein-based framework demonstrates superior performance, delivering both higher diagnostic precision and accelerated processing velocities across all three neurocognitive disorders.

Structural connectivity (SC) data are crucial for brain network analysis, but SC-based machine learning often suffers from limited data availability, hindering model generalization and robustness. Although data augmentation using deep generative models has attracted increasing attention, it remains unclear how different models capture the complex topological features of SC data. To clarify the learning characteristics of deep generative models for SC generation, this study compares three representative models: variational autoencoder (VAE), Wasserstein GAN with gradient penalty (WGAN-GP), and denoising diffusion probabilistic models (DDPM). We systematically evaluated these models using both synthetic datasets with known characteristics and real-world SC data. Generation quality was assessed using graph-theoretic metric comparisons and visual inspection of the generated adjacency matrices. WGAN-GP showed relatively stable performance across datasets and metrics, without severe performance degradation across evaluation settings. In contrast, VAE and DDPM performed well in specific aspects but were more sensitive to data characteristics. These findings suggest that WGAN-GP may serve as the most balanced baseline for future SC data augmentation studies, whereas VAE and DDPM may be useful depending on the target application and structural properties of interest. Furthermore, because all models struggled to fully reproduce strict global constraints such as planarity, our results suggest that standard generative models may be insufficient to capture the complex topological features of SC data. This highlights the importance of incorporating the desired structural properties into the training or generation process.

Computational models of brain structures impacted by Parkinsons disease are useful for exploring potential therapies. We use the Hindmarsh-Rose neuronal model to simulate synchronized activity in the subthalamic nucleus, capturing key features of the pathological rhythms observed in Parkinsons disease using a relatively small network of 100 neurons. Our model incorporates unidirectional excitatory chemical synapses whose strengths evolve according to a spike-timing-dependent plasticity (STDP) rule. To account for inputs from unmodelled neurons, both uniformly distributed white noise and Poisson noise were explored. White noise produced a single stable state of synchronized neuronal activity whereas Poisson noise resulted in two stable states, one synchronized and one desynchronized. We applied coordinated reset stimulation with a rapidly varying sequence (RVS CR) to examine its ability to reduce neuronal synchrony. The neuronal population was divided into subpopulations representing distinct physical sites of stimulation, as in deep brain stimulation therapy, and phase-shifted stimuli were delivered to each subpopulation in a random sequence. We explored how stimulation frequency and the number of stimulation sites affect the efficacy of RVS CR at desynchronizing the network. We demonstrate that RVS CR efficacy is sensitive to the depression-to-potentiation ratio in the STDP rule, which may be an important parameter to tune when reconciling simulations with experimental data. Numerical simulation of neuronal networks is constrained by computational resources when models demand large networks. This work proposes a model that demonstrates similar utility with a relatively small network, enabling researchers to study pathological neuronal activity and treatments more efficiently.

Approximately 50% of the population in India is estimated to experience sleep-related disorders. Sleep deprivation is a prevalent condition that adversely impacts cognitive performance, neural functioning, and overall health. Electroencephalography (EEG) offers an objective means of capturing neural alterations associated with sleep loss, making it well-suited for automated detection frameworks. In this study, we explore the application of a Quantum Support Vector Machine and Hybrid Quantum Neural Networks to classify sleep-deprived and well-rested states using resting-state EEG signals. A comprehensive feature extraction pipeline is employed, incorporating spectral band power, band ratios, Hjorth parameters, and functional connectivity measures. These features are subsequently encoded into quantum states to construct a quantum kernel, which is then utilized for classification. Model performance is evaluated under both epoch-level and subject-level data partitioning schemes. The Hybrid Quantum Neural Network (HQNN) achieves the highest performance across both evaluation settings, attaining an accuracy of 96.88% at the epoch level and 81.25% at the subject level. The QSVM model achieves accuracies of 93.75% and 75.00% for epoch-level and subject-level evaluations, respectively. At subject-level and epoch -level evaluation, HQNN outperforms previously reported results (68.23% and 95.72%). Overall, these findings highlight the potential of quantum machine learning as a competitive approach for EEG-based sleep deprivation detection, with promising implications for real-world biomedical applications.

Distinct smartphone interaction behaviors, like short-form video scrolling and mobile gaming, elicit qualitatively different cognitive and physiological responses. However, such distinctions is often overlooked by approaches that treat smart-phone use as a monolithic behavior. This paper presents Brain-Body Spatio-Temporal Transformer (BB-STT), a unified deep learning framework for classifying interaction-specific physiological signatures from multimodal signals, including EEG, EDA, PPG, and eye-tracking. BB-STT achieves 83.51% accuracy in distinguishing smartphone from non-smartphone activity and 74.13% accuracy in three-class classification of short-form video, gaming, and baseline viewing. The model demonstrates strong generalization with leave-one-subject-out (LOSO) performance that is also comparable to 5-fold cross-validation accuracy. Cross-modal attention emerges as the key component, improving three-class accuracy by 16.74 points through dynamic integration of multimodal signals. Interpretability analysis indicates a hierarchical organization of physiological responses. Eye-tracking features, particularly gaze depth, enable coarse separation between smartphone and non-smartphone activity. In contrast, finer discrimination between passive video viewing and active gaming on smartphones relies on the joint contribution of bilateral pupil dilation and central EEG features. Together, these results demonstrate the potential of multimodal physiological signals for objective, real-time assessment of digital engagement in naturalistic settings.

IntroductionTo leverage high-frequency oscillations (HFOs) as a biomarker with significant potential, this study compared a large set of detectors on a unified dataset, aiming to evaluate their clinical applicability under realistic conditions. MethodsEleven automatic detectors were applied to a retrospective dataset of intracranial and scalp EEGs from 27 consecutive pediatric patients. Inter-detector agreement was assessed using Spearmans Rho, and the area under the curve (AUC) for seizure onset zone (SOZ) prediction served as a consistent reference standard to enable reliable comparisons across recording modalities. Analyses were conducted separately for HFO and Spike-HFO detections. ResultsThe average age of our cohort was 12.4 years (SD 4.0; range 5-18). AUC values in scalp EEG ranged from 0.61 to 0.67 for HFOs and from 0.53 to 0.63 for Spike-HFO. AUC values in intracranial EEG ranged from 0.48 to 0.66 for HFOs and 0.54 to 0.69 in Spike-HFO. Although only three of the 11 detectors were specifically developed or adapted for scalp EEG, the detectors generally achieved higher AUC values and stronger agreement in scalp EEG ConclusionsWe present the first study comparing intracranial and scalp detectors by testing them beyond the modalities for which they were originally designed. Although the clinical utility of detections was comparable across EEG modalities, it remained lower than reported in original studies assessing the diagnostic value of HFOs. Caution is warranted when applying a publicly available detector to a new dataset, and detector robustness remains a critical issue. Key points- A comprehensive head-to-head comparison of 11 detectors demonstrated significant variability in detector agreement and clinical utility - Clinical utility was not necessarily linked to the EEG recording type the detector was originally designed for - Despite widely accepted use of automatic detections, detector robustness remains a critical issue

Background: Free-text EEG reports typically lack structure, hindering scalable analysis. We evaluate a large language model (LLM) pipeline to extract structured diagnostic labels and confidence levels from these reports. Methods: We developed a hierarchical annotation schema to classify EEG reports for four specific abnormality types using a four-point confidence scale. To establish ground truth, two certified EEG technicians annotated a diverse dataset of reports authored by neurologists with distinct writing styles. We then implemented a grammar-constrained Mistral-7B pipeline, iteratively prompt-tuned on a development set to mirror these expert annotations. The pipeline's effectiveness was evaluated against the human expert benchmark using core agreement (diagnostic accuracy) and certainty-adjusted agreement (confidence alignment), with classical NLP models serving as a secondary baseline. Results: Mistral-7B significantly outperformed baselines, achieving 96% accuracy for overall abnormality detection, approaching the human benchmark of 98%. Crucially, the model successfully identified rare epileptiform abnormalities where traditional models failed and generalized robustly across distinct reporting styles. While diagnostic accuracy was high, a performance gap persisted in certainty-adjusted agreement, indicating that accurately modeling nuanced clinical confidence remains a challenge. Conclusion: LLMs can effectively automate the extraction of structured diagnostic information from EEG reports with near-human accuracy and strong generalization. While confidence calibration requires further refinement, the combination of accurate classification and explainability makes this pipeline a promising tool for standardizing clinical data at scale. Keywords: Routine Clinical Electroencephalography; Large Language Models; Clinical NLP; Confidence Assessment; Explainable AI; Neurophysiological Evaluation

Major depressive disorder (MDD) is a common psychiatric disorder that requires reliable and objective assessment for early clinical intervention. Electroencephalography (EEG) is widely used for this purpose because it provides a non-invasive and low-cost measure of brain activity with high temporal resolution. However, EEG-based depression detection remains challenging due to the nonlinear nature of EEG signals, inter-subject variability, and the limited availability of subject-independent evaluation. To address these issues, this paper proposes a hybrid quantum-classical multiscale long short-term memory with parameterized quantum circuit branches (MS-LSTM-PQC) framework for subject-level EEG-based depression detection. The proposed model extracts temporal representations at multiple scales using parallel LSTM branches and incorporates eyes-closed (EC) and eyes-open (EO) condition information through condition-aware feature fusion. To further enhance the learned representations, scale-specific LSTM features are processed using PQC-based quantum branches implemented with TensorFlow Quantum (TFQ), providing an additional nonlinear feature transformation before classification. Experiments were conducted on the Mumtaz EEG depression dataset using EC-only, EO-only, and merged EC+EO conditions with 1-s, 2-s, and 3-s EEG windows. To reduce subject-level data leakage, all experiments were evaluated using 5-fold and 10-fold GroupKFold validation. The best overall accuracies across the evaluated settings were 92.05% and 95.08% under 5-fold and 10-fold GroupKFold validation, respectively. The 2-s merged EC+EO setting provided the most stable performance across validation protocols. In addition, Integrated Gradients (IG)-based explainability analysis showed that frontal and fronto-central channels, especially Fz, showed higher contributions to the model decision. These results suggest that multiscale temporal learning with quantum-enhanced feature transformation can support subject-level EEG-based depression detection under leakage-controlled evaluation.